Propylene Glycol

[WK]

Dear All,
I have some very poor reproducibility with determining PG in methanol as a sample solvent. I use isobutyl propionate as internal standard. I use "5" type phase, autosampler, FID, split.Have tried various split flows and injector temperatures. I replaced wool packed liner and this improved for about 10 runs then it went back to poor state again - ie consecutive runs 65% and 72%. I have searched the forum for info but nothing helpful there.
Any ideas?
WK

[Peter Apps]

Almost certainly you have adsorption in the inlet. You can try increasing the inlet temperature, which might give a few more runs before you have to change liners. It will also help if you use a glycol rather than an ester as the internal standard.

Peter

[Consumer Products Guy]

Normally I derivatize glycols with BSTFA and use a "5" type column, like HP-5. what levels of propylene glycol are you looking at, and is methanol just your dilution solvent or your sample? We assay for propylene glycol routinely in both anionic raw materials and finished products such as transparent bar soaps after trimethylsilyl derivatization. Of course, methanol will also react with that, that's why I wanted to know whether you can use a different solvent, or the methanol could be OK if it's your actual sample, as long as you use excess reagent.

For underivatized propylene glycol, I might suggest a -624 column, like Rtx-624, using conditions like for EG and DEG in the USP glycerin monograph, which we do run.

[WK]

Thanks Peter and CPG,
I will try running underivatized PG on my 624 today - I will keep you posted. You are right - I should use a glycol as internal standard (the PG peak varied - the internal standard area consistent)
WK

[WK]

Peter and CPG,
I have tried same on 624 with DEG as internal standard and methanol solvent but with same effect.
4 or 5 runs ratio 1.80 to 1.82 then next run 1.9 and variable then on.
That would appear to be the adsorption type effect - Peter?
I am using wool packed liner at 240degC.
I noticed yesterday that at injector 160degC the effect was much worse.
Arrgghhh! Ha-Ha! Doohh! Grrhhh!
WK

[Consumer Products Guy]

We first used water as solvent (as per USP glycerin monograph) then changed to methanol solvent with -624 column when USP updated the monograph. Remember, we were looking for EG and DEG, not PG there. That said, 99.9% of the time here that we assay for PG, we DO derivatize, then assay on a DB-1 or DB-5 type column. We routinely use external standard quantitation, and meet system suitabilty RSD <2%. We have used internal standard for one of our facilities, and we used butanediol for that, but DEG or glycerin should be fine. Can you disclose exactly what your "sample" is, and at approximately what levels your are trying to quantitate? It doesn't sound to me like your real question is to determine PG in methanol, that it's to determine PG in a different type of sample.

[WK]

Yes CPG,
You are right - I am only using methanol as a solvent. I need to determine PG in flavours and extracts.
Anything between 20-90% and rest can be water,ethanol, glycerin, plant materials, plant pigments etc.
Please could you suggest a suitable solvent when using derivatization reagent?
(My RSD with outliers is approx. 2.2% and without is 0.6%)
WK

[Peter Apps]

The glass wool is likely to be bad news, a huge area that can become adsorptively active. Try a liner with no wool, or just the tiniest bit to help sample vaporization. Ultimately you might need a baffled liner. Changing injection parameters such as pre-injection dwell might also help a bit, methanol does not like the fast cold needle injections that are default on a certain well known brand of instrument.

Peter

[Consumer Products Guy]

Yes CPG,
You are right - I am only using methanol as a solvent. I need to determine PG in flavours and extracts.
Anything between 20-90% and rest can be water,ethanol, glycerin, plant materials, plant pigments etc.
Please could you suggest a suitable solvent when using derivatization reagent?
(My RSD with outliers is approx. 2.2% and without is 0.6%)
WK

We normally assay samples of 0 - 10% PG. We make up sample solutions about 2% in DMF (N,N-dimethylformamide), then mix (in an autosampler vial) 2 parts of that with 1 part of BSTFA containing 1% TMCS (trimethylchlorosilane), cap, shake, and inject. There was a publication in JAOCS in the early 1980s doing similar for glycerin. This is very easy and reproducible. Run isothermal on GC anywhere from 80 to 120C, then program to get the other stuff off the column. Inject a DMF-BSTFA blank as well to clean out any old crap off the inlet and column, and so you can determine PG-TMS retention time; it will elute just after the DMF-BSTFA does as it's pretty volatile.

[WK]

CPG & Peter,
Thanks again for the posts - I was hoping to just inject - but you have convinced me this is not the way to go.
I will try to develop a derivatization method based on CPG method.
CPG - would I get good resolution running derivatized samples on a 5 phase or is a 624 necessary?
Regards
WK

[Consumer Products Guy]

I will try to develop a derivatization method based on CPG method.
CPG - would I get good resolution running derivatized samples on a 5 phase or is a 624 necessary?
Regards
WK

Don't use the 624 phase for this, I've never tried; apparently it's 6% cyanopropylphenyl/94% dimethyl polysiloxane, so it doesn't sound like it'd be damaged by BSTFA, though, I've just never tried. 624 is a polar phase, and derivatizing the PG makes it non-polar.
Use the HP-5, DB-5 type column for this assay, or even an HP-1/DB-1. Use the 624 for stuff that's polar.

[Don_Hilton]

If it is of any help, I used to work with an analysis for propylene glycol and some other items in tobacco extracts. We used a PEG type column. The PEG type column works well with PG, as opposed to the much more non-polar columns. A quick look on the net showed a couple of links (http://www.hc-sc.gc.ca/hl-vs/tobac-taba ... nt-eng.php or http://www.coresta.org/Recommended_Methods/CRM_60.pdf) that look to be pretty close to what I used to do. I did not notice an inlet liner in the articles. We always used liners with glass wool plugs. They comment on using deactiveated liners - and we discovered that deactivated wool is imortant - with deactivation being done on liners after the wool is inserted in the liner. (Otherwise you have deactivated wool with many very active broken ends - formed during the process of packing the liner.)

We made standards in matrix and ran several old samples from the day before down the column before beginning analytical work - to coat active sites, as activity can be a real problem. The coresta method says to use standards for conditioning columns, but you can use yesterday's samples while getting today's standards ready to run.